The short answer
There is no single "correct" QR code size. The minimum size depends entirely on how far away the scanner will be from the code. A QR code on a table card needs to be around 25–40 mm. A QR code on a billboard needs to be 1,000 mm or more.
The rule of thumb: minimum QR code width = scan distance divided by 10.
At 30 cm scan distance: minimum 30 mm wide. At 100 cm: minimum 100 mm wide. At 2 metres: minimum 200 mm wide.
This is a safe minimum. Going larger always improves reliability. Going smaller increases the chance of scan failure, particularly on older phones or in poor light.
Why size matters more than resolution
A common misconception is that scan failures are caused by low resolution - the image is blurry or pixelated. In practice, most scan failures are caused by the code being physically too small for the distance it is scanned from.
Phone cameras decode QR codes by reading the ratio of light and dark modules (the small squares that make up the code). If the code is too small relative to the camera's field of view at the intended distance, the camera cannot distinguish individual modules. Printing at higher resolution on the same physical size does not fix this.
What is the quiet zone and why does it matter?
The quiet zone is the white border around the QR code. It is not decoration - it is a functional part of the standard. Without a sufficient quiet zone, the decoding algorithm cannot determine where the code starts and ends.
The minimum quiet zone is 4 module widths on each side. A module is one small square in the QR grid. If you crop the image tightly in your design software, you may be cutting into this zone.
If a QR code scans fine on screen but fails after printing, a cropped quiet zone is the most likely cause - it is also one of the hardest things to see at a glance. Use the QR quiet zone checker to confirm the margin is sufficient before going to print.
How many modules does a QR code have?
The number of modules depends on the data capacity and error correction level. A minimal URL with high error correction might use a 29×29 grid. A vCard with a lot of data might use a 41×41 or larger grid.
More modules at the same printed size means smaller individual modules, which are harder for a camera to resolve. If you need to fit a lot of data into a small space, lower the error correction level and keep the content as short as possible.
Error correction and size
The four error correction levels - L, M, Q, H - affect both the resilience and the density of the code:
- L (7% recovery): Fewest modules, largest individual squares, easiest to scan at small sizes. Use when the code will be in a clean, controlled environment.
- M (15%): Good default for general use.
- Q (25%): Use when there is some risk of partial obscuring - for example, a logo overlay or outdoor exposure.
- H (30%): Use for codes in challenging environments or those carrying a logo. Higher recovery capacity comes at the cost of more modules and therefore smaller squares at the same print size.
For a small code on a table card, L or M may scan more reliably than H at the same size, because the modules are larger and easier for the camera to read.
Checking your size before printing
Use the QR print size checker on this site to calculate the minimum safe size for any intended scan distance. Enter your code's module count and the maximum scan distance, and the tool tells you the minimum print width and whether your intended size is likely to scan reliably.
The QR contrast checker is also worth running if you are using anything other than black on white - colour combinations that look distinct on screen can fail the contrast threshold that QR decoders rely on.
Practical size recommendations
| Use case | Typical scan distance | Minimum width | |---|---|---| | Table card / menu | 25–50 cm | 30 mm | | Business card | 15–25 cm | 20 mm | | Poster / A4 flyer | 50–100 cm | 60 mm | | Retail shelf label | 15–30 cm | 25 mm | | Shop window | 50–200 cm | 80–150 mm | | Outdoor banner | 1–5 m | 150–400 mm |
Always test-scan a printed proof at the intended distance before committing to a full print run. Screen previews are not a reliable substitute for a printed test.